About
Data platforms, and the tools people use on top of them. Based in the UK.
Now
- Data engineering. Moving teams onto the new metric engine, ahead of its full launch.
- AI. Getting the KPI assistant ready for its first release.
- Agentic workflows. AI agents that write and review code, raise tickets, update docs and prepare releases.
- Automation. A browser tool, driven by Jira tickets, moves dashboard content between environments. A team app syncs SharePoint bug lists into Jira.
Background
Nine years in data and BI. The work began at a consultancy, with reports and dashboards for commercial teams in Qlik and Power BI. Many of those reports were for senior leaders.
Then came data engineering: PySpark and SQL pipelines over several regional CRM systems, finance data and market research. The transforms clean and join the data, so every report reads one consistent model.
Next came platform work. One reporting team cannot build every report. So the work moved to the back end of a self-service BI tool. There, people build their own dashboards from one set of governed KPIs.
Today: tech lead for a team of up to seven data, UI and AI engineers. The work covers the shared data model and its protection, platform cost, agentic workflows and automation. Evenings go on personal projects, such as Ward, Rook and Trellis.
What the work covers
AI agents
Agents that find duplicate metrics, review code changes and write code for review.
Agentic workflows
One rule set and skill library for every agent. Also on personal projects: Rook and Trellis.
Automation
Jira-driven tools that move dashboards and sync SharePoint bug lists.
Pipelines and transforms
PySpark and SQL pipelines that bring sales, finance and research data into one model.
Back-end APIs
The metric engine and the APIs that screens and AI agents call.
Data model and privacy
One shared data model, with private columns protected in the model, not in each app.
Row access
Each person sees the records that their place in the sales structure allows.
Security reviews
Platform reviews with the security team, and closing the access risks they find.
Platform operations
Data connections, network rules, app registration, and access and change reviews.
Infrastructure
Docker, Kubernetes and Helm on Trellis, a personal project, with TLS and backups.
Cost
Mapping the AWS bill to the teams that drive it, then cutting the platform usage behind it.
Testing and releases
Regression suites that compare old and new answers before any release.
Away from the screen
Photography
Mostly landscapes: low sun, still water, open fields.
Hiking
Hills in every kind of weather, from clear ridges to frozen summits.
Camping
Nights outdoors, by an open fire.
Contact
Have a question about a post? Want to compare notes on data platforms and AI cost? Send an email.